Beyond model interpretability using LDA and decision trees for α‐amylase and α‐glucosidase inhibitor classification studies

Abstract

In this report are used two data sets involving the main antidiabetic enzyme targets α‐amylase and α‐glucosidase. The prediction of α‐amylase and α‐glucosidase inhibitory activity as antidiabetic is carried out using LDA and classification trees (CT). A large data set of 640 compounds for α‐amylase and 1546 compounds in the case of α‐glucosidase are selected to develop the tree model. In the case of CT‐J48 have the better classification model performances for both targets with values above 80%–90% for the training and prediction sets, correspondingly. The best model shows an accuracy higher than 95% for training set; the model was also validated using 10‐fold cross‐validation procedure and through a test set achieving accuracy values of 85.32% and 86.80%, correspondingly. Additionally, the obtained model is compared with other approaches previously published in the international literature showing better results. Finally, we can say that the present results provided a double‐target approach for increasing the estimation of antidiabetic chemicals identification aimed by double‐way workflow in virtual screening pipelines.

In Silico Approaches to Identify Polyphenol Compounds as α-Glucosidase and α-Amylase Inhibitors against Type-II Diabetes

“…In vitro and in vivo tests for T2DM can also evaluate the toxicity of drugs or compounds in development and the modified activity, and bioavailability of herbal medicine compounds [ 280 ]. Importantly, preliminary study includes protein-ligand interactions comparable to the lock and key principle [ 281 ]. The major potent force for binding is hydrophobic interaction, while in silico modeling can be helpful to identify drug target via bioinformatics tools [ 282 ].…”

Tree-Based QSAR Model for Drug Repurposing in the Discovery of New Antibacterial Compounds against Escherichia coli

“…Regarding the methodology used to develop de tree-based QSAR model, LDA is a frequently used method for the development of QSAR models [25][26][27][28], however, it has a key limitation: variables with discrete values cannot be used for the development of the model since they cannot be analyzed using the same statistical methods [29]. This is an important issue in the development of QSAR models because discrete descriptors contain relevant information that could be useful in predicting pharmacological activity.…”

“…This is an important issue in the development of QSAR models because discrete descriptors contain relevant information that could be useful in predicting pharmacological activity. The formalism for the development of QSAR/QSPR models using LDA is based on the hypothesis that there is a group of compounds, (Ω), where each compound shall be denoted by ω ∈ Ω and the existence of a series of indexes we will design as {I i } i∈I , each element of this set is formed by the function I : Ω → R , so that each compound has an assigned value for each index described [28]. If I(Ω) takes values in a discrete or numerable set, we face a discrete index.…”

Towards Dual Antibacterial Drug-Nanoparticle (DADNP) Systems Computational Design

“…IFPTML models training and validation. In the first instance, the Linear Discriminant Analysis (LDA) approach was employed to find the basic model [18]. To choose the input characteristics automatically, we employed the Forward Step-Wise (FSW) technique as a variable selection strategy.…”